Resistor thermal issues in a SWR Studio 200

[wizard333]

I have a SWR Studio 200 amp, just replaced the tube and electrolytics.

In the power section there are 2 dropping resistors right after the big power filter caps that drop the ~ +/- 60v down to the +/- 15v supply. Schematic calls for 3W 660 ohm, what is present are a 600 ohm and 100 ohm in series. Looks like a 2W 600 ohm metal film or metal oxide, and a 1W carbon film 100 ohm. They are from the factory that way.

They are putting out enough heat to cook an egg on, and they sit right between the 2 big filter caps. The Caps are getting heated up from that, and since those caps went bad before, I'd expect that heat to shorten their life again.

Any suggestions for cooling this situation off with a different resistor type etc?

[Enzo]

Sure. Bigger resistors. get a pair of those aluminum extruded things with little mounting flanges. Screw them to the chassis and run wires to the board.

[wizard333]

That's a good idea, except I can't find anything close to that ohm rating in that type of resistor at Mouser or Digikey. Any idea on a source?

[Enzo]

Nothing close? 660 is an odd value, but 680 is a standard value. The circuit is obviously not critical, besides 680 is within 5% of 660.

A plain old 5w 680 would probably work for you. Mouser 280-CR5-680-RC

But in looking for the chassis mount akuminum extruded guys, 680 ohm 10w, 10w being the smallest aluminum ones, the Arcol for $2.31 is in stock, item # 284-HS10-680F. And the Vishay/Dale also in stock for $2.83, item # 71-RH10-680.


You WON'T be burning those up.

[wizard333]

When I looked, Mouser only had one of the 5W 700s in stock, I'd need two; and no 680s in stock. The "in stock" thing is usually the killer with Mouser. Didn't look at the 10W though. I'll have to measure and ponder where I can mount them, but that sounds like a plan if they'll fit.

Even if the stated 3W 660 was available, the calculated dissipation is very close to 3W, slightly over actually, so no head room (45v drop across 660 ohms). I figured 5W would be plenty of headroom, 10W should be pure overkill.

Thanks Enzo!

[MarkusBass]

This is typical problem with SWR amps. I usually put two 330 Ohms/5W ceramic resistors in series (x2). It means that I get two 660 Ohms /10W resistors. No more problems.

Mark

[Enzo]

Mouser currently has 468 of the 5 watt 680 ohm cement resistors in stock at 44 cents. - stock 280-CR5-680-RC

They have 7 different 680 ohm 5 watt resistors in stock, the least of any is 80 pieces, the most was 1071


10w may be overkill, but it gets you a chassis mountable part, so it gets you off the board.


Remember, the circuit will dissipate the 3 watts no matter what. Larger wattage resistors will just not be as intense a hot spot, the heat will be spread. Chassis mount will conduct the heat into the chassis and run much cooler.

[R.G.]

This appears to be one of those things where MBA Disease is simply not going to let the amps get better. This problem happens in many amps. The amp maker is too cheap to specify a transformer winding to make +/-15Vdc for opamps, and so they do it by putting a couple of zener diodes and some power resistors to drop the voltage from the main power supply to +/-15V. This is done by designers who have never had to live with the results of their actions, and encouraged by MBAs who are clueless and simply on their way to some better job on the ladder.

The fundamental problem is that zener diodes are a modestly unreliable shunt voltage regulator unless you're very, very good with designing it. They waste a lot of power, get hot, heat other things, etc. A 7815 + 7915 regulator pair would fix this for good, as they contain quite a bit of smarts as regulators. But the amps these are used in have power supply voltages over the 40V maximum that the 7800/7900 series can stand. So the "designers" fall back on power resistors and zeners.

An overcurrent on a resistor limited shunt regulator overheats the resistor. Most of these amps get into this at some point in their life. This gets the minimally rated resistors hot enough to unsolder themselves and char the PCB in some cases.

The cheap solution is to put a really high power resistor in there to spread the heat, as noted, so the point temperature stays below the melting point of solder and the charring point of glass epoxy. The good/reliable solution is to do some thing with discrete parts to convert it from a shunt regulator to a series regulator. The LM317/337 regulator pair plus a power transistor and a few other parts steps around this trap pretty well.

But most people can't cope with that level of complexity. So get some chassis mount power resistors, as noted. I really like the thick film power resistors in a two-lead TO-220 package. When properly heat-sunk, these will safely dissipate 20W and up. They're semi-expensive, but boy are they slick. Mouser has 20W 680R resistors in a TO-220 package for under $5.00 each.